Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Jan 20.
Published in final edited form as: J Biomech. 2024 Jan 20;163:111960. doi: 10.1016/j.jbiomech.2024.111960

Association of quantitative diffusion tensor imaging measures with time to return to sport and reinjury incidence following acute hamstring strain injury

Christa M Wille 1,2,3, Samuel A Hurley 4, Mikel R Joachim 1,3, Kenneth Lee 4, Richard Kijowski 5, Bryan C Heiderscheit 1,2,3
PMCID: PMC10923138  NIHMSID: NIHMS1963631  PMID: 38290304

Abstract

Hamstring strain injuries (HSI) are a common occurrence in athletics and complicated by limited prognostic indicators and high rates of reinjury. Assessment of injury characteristics at the time of injury (TOI) may be used to manage athlete expectations for time to return to sport (RTS) and mitigate reinjury risk. Magnetic resonance imaging (MRI) is routinely used in soft tissue injury management, but its prognostic value for HSI is widely debated. Recent advancements in musculoskeletal MRI, such as diffusion tensor imaging (DTI), have allowed for quantitative measures of muscle microstructure assessment. The purpose of this study was to determine the association of TOI MRI-based measures, including the British Athletic Muscle Injury Classification (BAMIC) system, edema volume, and DTI metrics, with time to RTS and reinjury incidence. Negative binomial regressions and generalized estimating equations were used to determine relationships between imaging measures and time to RTS and reinjury, respectively. Twenty-six index injuries were observed, with five recorded reinjuries. A significant association was not detected between BAMIC score and edema volume at TOI with days to RTS (p-values ≥ 0.15) or reinjury (p-values ≥ 0.13). Similarly, a significant association between DTI metrics and days to RTS was not detected (p-values ≥ 0.11). Although diffusivity metrics are expected to increase following injury, decreased values were observed in those who reinjured (mean diffusivity, p = 0.016; radial diffusivity, p = 0.02; principal effective diffusivity eigenvalues, p-values = 0.007-0.057). Additional work to further understand the directional relationship observed between DTI metrics and reinjury status and the influence of external factors is warranted.

Keywords: Diffusion Tensor Imaging, Skeletal Muscle, Hamstring Strain Injury, Return to Sport, Reinjury

Introduction

Muscle strain injuries, specifically of the hamstrings, are a common occurrence in athletics and complicated by limited prognostic indicators and high rates of reinjury (de Visser et al., 2012; Opar et al., 2012). Hamstring strain injuries (HSI) may result in significant loss of time from activity, decreased quality of life (LIu H, 2012), and increased financial burden in elite sports (Hickey et al., 2014). The use of injury characteristics at the time of injury (TOI) has considerable value in managing athlete expectations, guiding activity progression, and mitigating reinjury risk, yet clear associations are lacking between injury characteristics at TOI and clinical outcomes such as time needed to return to sport (RTS) and reinjury.

Magnetic resonance imaging (MRI) is used to aid in soft tissue injury management and prognosis in a sports medicine setting (Guermazi et al., 2017). MRI-based injury grading scales, such as the British Athletics Muscle Injury Classification (BAMIC), have been associated with time to RTS (Biglands et al., 2020; McAleer et al., 2022; Pollock et al., 2016; Shamji et al., 2021; Tears et al., 2022). However, the prognostic value of MRI-based assessments within days following an acute HSI is widely debated (Greenky and Cohen, 2017; Reurink et al., 2015; Reurink et al., 2014; Rudisill et al., 2021; Wangensteen et al., 2015), and MRI-based assessments do not appear to improve associations with time to RTS beyond physical examination metrics and athlete HSI history (Jacobsen et al., 2016; Wangensteen et al., 2015). Further, the majority of these assessments are at the gross anatomical level (i.e., volume, location, number of muscles involved) (Rudisill et al., 2021). The conflicting evidence surrounding the potential association of MRI measures and time to RTS is further complicated by the high variability in RTS time observed following HSI and further strengthens the need for improvements in the tools available to inform prognosis following HSI.

Results measuring the association of reinjury and TOI MRI assessment following HSI are equally mixed. Some findings suggest reinjury risk is associated with the hamstring muscle or the tissue-type involved in the index injury (van Heumen et al., 2017), while others have found that MRI descriptors of the index HSI are not associated with risk of reinjury (Green et al., 2020).

Recent advancements in musculoskeletal MRI such as diffusion tensor imaging (DTI) have allowed for additional quantitative measures of muscle microstructure at the cellular level. DTI generates contrast from the random diffusion of water molecules, and utilizes differences in diffusivity to infer information about muscle microstructure. The microstructure of an intact muscle fiber will encourage water to preferentially diffuse along the direction of the fiber, while diffusion will occur in a more random and directionally isotropic pattern in injured or damaged muscle fibers (Damon et al., 2017). Preliminary evidence demonstrates that DTI is sensitive to detect differences in the injured limb following acute HSI (Monte et al., 2023). However, the association of DTI parameters following acute HSI and longitudinal clinical outcomes such as time to RTS and reinjury incidence has not yet been explored. Therefore, the purpose of this study was to determine the association of MRI-based injury grading and quantitative imaging measures at the time of acute HSI with time to RTS and reinjury.

Methods

Study Design

The data presented in this study were collected as part of a larger prospective cohort investigation of collegiate athletes who sustained an HSI. The study was approved by the University’s Health Sciences Institutional Review Board, and participants provided written informed consent prior to enrollment.

Participants

Collegiate football, soccer, and track athletes who sustained a unilateral HSI confirmed by a member of the respective teams’ sports medicine staff were included in this study. An HSI was diagnosed as sudden onset of posterior thigh pain that occurred during a sport-related activity that resulted in the athlete not being able to return for at least one practice or competition, and the presence of two or more of the following during clinical examination: palpable pain along the hamstring muscles, posterior thigh pain without radicular symptoms during a passive straight leg raise, and/or weakness or pain with resisted knee flexion (Heiderscheit et al., 2010). Participants were excluded from this analysis if imaging data was not captured within seven days of the HSI or if there was not an imaging confirmed HSI identified by a musculoskeletal radiologist (KL, RK). The presence of injury on T2-weighted MR was identified as a region of architectural disruption of the hamstring muscle complex and/or hyperintense signal within the hamstrings, most likely representing injury associated edema. Athletes with distinct injuries of each limb that occurred at different timepoints within the study observation window were allowed to be included in the analysis, however when an athlete sustained multiple injuries of the same limb within the observation window, only imaging data from the first HSI was included in the analysis.

Injury Management

A standardized rehabilitation protocol was implemented by the teams’ athletic trainer. In brief, the rehabilitation guidelines followed three phases with defined goals: Phase 1.) Protect scar development, minimize atrophy and emphasis core, isometrics, and mid-range isotonics; Phase 2.) Regain pain-free hamstring strength at mid-range and progress to longer hamstring length, develop neuromuscular control of trunk and pelvis with progressive increase in movement speed; Phase 3.) Symptom free during all activities, normal concentric and eccentric hamstring strength through full range of motion and speeds, emphasis on eccentrics and sport specific drills (Sherry et al., 2015). RTS was determined when medical clearance was obtained to resume all sport-related activities. RTS clearance was based on a combination of factors including full hamstring range of motion, minimal to no pain with hamstring palpation, and no apprehension with on-field sports-specific movements. Athletes and their teams’ respective sports medicine and coaching staff were blinded from the use of imaging data and analyses; thus, any imaging data acquired for the purpose of this study was not used to guide rehabilitation or RTS decision making. Following RTS, all athlete reinjuries were tracked by the team athletic trainers. Reinjuries were defined as an acute HSI to the same limb as the index HSI, requiring the athlete to miss at least one practice or competition within a 12-month period after RTS. Any HSIs involving the contralateral limb, relative to the index HSI, were not included as a reinjury.

Magnetic Resonance Imaging Protocol

Participants received an MRI examination of bilateral upper thighs, completed on a 3.0T scanner (GE Healthcare Discovery MR750, Waukesha, WI) using a 32-channel full torso coil and positioned in a feet-first supine position in the scanner. A 3D axial T1-weighted spoiled gradient recalled echo (SPGR) sequence was used for anatomical reference with the following parameters: no fat saturation, TR/TE = 5.9/2.1 ms, flip =15 degrees, FOV = 44 cm, matrix = 640x640 (reconstructed at 1024x1024), 80 slices, 5 mm thick, bandwidth = 195 Hz/pixel, PURE intensity correction. A T2-weighted fast recovery fast spin echo (FR-FSE) sequence was used to identify muscle edema in the region of injury with the following parameters: fat/water separation with iterative decomposition of water and fat with echo asymmetry and least-squares estimation (IDEAL), TR/TE = 4,473/85.0 ms, matrix = 448 x 448 (reconstructed to 512 x 512), 44 slices, 7 mm thick, 2 mm spacing, bandwidth = 140 Hz/pixel. Diffusion-weighted images were acquired in two slabs using spin-echo echo planer imaging (SE-EPI) with weighting in 30 uniformly distributed directions on a unit sphere, 6 non-diffusion-weighted volumes (b=0 images), b-value of 500 s/mm2, and other parameters: TR/TE = 5770/51.1 ms, FOV = 48 cm, matrix = 160x160, 72 slices, 3 mm thick. The diffusion acquisition was repeated twice with reversed phase-encode directions (anterior-posterior and posterior-anterior) to correct for susceptibility-induced distortions.

Magnetic Resonance Imaging Analysis

Clinical interpretations of the HSI were performed by one of two musculoskeletal radiologists (RK, KL), each with over 20 years of experience. The primary muscle of injury was identified and the location and severity of injury was evaluated using the BAMIC (Pollock et al., 2014) scoring system, which assesses the overall injury grade (0-4) and site classification (myofascial [a], musculotendinous [b], intratendinous [c]), and was assessed using the T1- and T2-weighted sequences.

For diffusion-weighted data, distortion, eddy current, and motion correction were performed using FSL TOPUP (Andersson et al., 2003) and EDDY (Andersson and Sotiropoulos, 2016) (Jenkinson et al., 2012). Data were filtered using a local principal component analysis filter (Manjon et al., 2013), then linear fitting to a diffusion tensor (DTI) model was performed with FMRIB Diffusion Toolbox (FDT, FMRIB Software Library, Oxford, UK). When small misalignments between T1- and diffusion-weighted scans occurred, they were manually registered with visual confirmation. Quantitative scalar measures were computed from DTI images, including fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD), and principal effective diffusivity eigenvalues (λ1, λ2, λ3) and were calculated using FMRIB’s Diffusion Toolbox (FMRIB Software Library, Oxford, UK) (Jenkinson et al., 2012).

Anatomical three-dimensional contours of each hamstring muscle (biceps femoris short head, biceps femoris long head, semitendinosus, semimembranosus) of bilateral thighs were completed via manual segmentation using T1-weighted axial SPGR images (FSLeyes, v1.5.0, Oxford, UK) (McCarthy, 2022). Using the T2-weighted image, musculoskeletal radiologists (KL, RK) identified the region of increased signal that was associated with an HSI on the injured limb. Manual segmentation was used to identify all voxels within this region (Figure 1) (FSLeyes, v1.5.0, Oxford, UK) (McCarthy, 2022) and was further refined by taking the intersection of the muscle boundaries and the representative region. The resulting region was used to represent the region of injured muscle tissue. Edema volume within the hamstring muscles was calculated by multiplying voxels identified as containing edema within the muscle boundaries by voxel volume. To account for differences in subject body habitus, volumes were normalized by the height-mass product (Handsfield et al., 2014).

Figure 1.

Figure 1.

Open in a new tab

Mean quantitative diffusion metrics were calculated within manually outlined regions of interest (ROI) on the injured limb defined by the intersection of A.) hamstring muscle boundaries on a T1-weighted image and B.) hyperintense signal, most likely representing injury associated edema on a T2-weighted image. Data shown are from one slice of a representative participant.

Biceps femoris short head (BFsh), biceps femoris long head (BFlh), semitendinosus (ST), semimembranosus (SM).

Mask data were down-sampled to match the resolution of diffusion-weighted imaging sequences and the resulting masks were used for the region of injury. Masks representing injury regions were superimposed over the FA, MD, RD, and principal effective diffusivity eigenvalue (λ1, λ2, λ3) maps to measure the mean values of quantitative DTI outcome measures within the region of injured muscle tissue (MATLAB, v2021b, MathWorks, Natick, MA).

Statistical Analysis

Standard descriptive statistics (means/standard deviations, median/interquartile range, and frequencies/percentages) were used to describe the participants. Imaging outcome measures included: edema volume within the hamstring muscles normalized by height*mass product; BAMIC grade and site classification; and DTI metrics (FA, MD, RD, λ1, λ2, λ3). The associations between imaging outcomes and time to RTS were modeled using a negative binomial regression, and reported as incident risk ratios (IRR) and 95% confidence intervals (CI). The associations between imaging outcomes and reinjury status were determined using generalized estimating equations (GEE) for a binomial outcome with a log link, and reported as odds ratios (OR) and 95% CI. All analyses were conducted using SAS v9.4 (SAS Institutes, Cary, NC) and significance was assessed at α ≤ 0.05.

Results

Twenty-two unique athletes met eligibility criteria and participated in this study, with four athletes sustaining two distinct injuries, one of each limb, resulting in 26 recorded HSIs (Figure 2). Participant characteristics at TOI are presented in Table 1. The median days to RTS was 22 (interquartile range 15-33.5); five reinjuries were recorded (19%) (Table 1) with the median time to reinjury being 47 (interquartile range 32 – 85) days after RTS of the index injury. The biceps femoris long head was the most commonly injured muscle (73%) with BAMIC classification 3c being the most common injury grade and location (31%). Imaging metrics for all HSIs are presented in Table 2. Representative imaging data for one participant is demonstrated in Figure 3.

Figure 2.

Figure 2.

Open in a new tab

Participant inclusion criteria. All participants included in this analysis had unilateral evidence of injury on a T2-weighted magnetic resonance image (MRI) within 7 days of injury. Participants who demonstrated evidence of bilateral edema or sustained bilateral injuries linked to the same date/mechanism of injury were excluded from the analysis. If a participant sustained distinct injuries on each limb on different dates, both injuries were included in the analysis, however when an athlete sustained repeat ipsilateral injuries within the study observation window, only imaging data from the first injury on that limb was included in the analysis.

Table 1.

Participant characteristics. Values are reported as counts and means (standard deviations) unless otherwise noted. Percentages are relative to the number of participants within each group summarized in each respective column. For participants with multiple injuries, descriptive statistics are reported for the first injury observed.

Reinjury
All (n = 22) Yes (n = 5) No (n = 17)
Age (years) 19.8 (1.3) 20.2 (1.9) 19.7 (1.9)
Height (m) 1.84 (0.07) 1.86 (0.02) 1.83 (0.02)
Weight (kg) 88.4 (18.4) 87.1 (9.9) 85.4 (9.9)
Sex (% of total participants)
 Male 19 (86%) 5 (100%) 14 (82%)
 Female 3 (14%) 0 (0%) 3 (18%)
Sport (% of total participants)
 Football 10 (45%) 1 (20%) 9 (53%)
 Track 11 (50%) 4 (80%) 7 (41%)
 Soccer 1 (5%) 0 (0%) 1 (6%)
Time to Return to Sport (days) (median, (IQR)) 22 (15-33.5) 21 (19-22) 22 (12.5-45)
Open in a new tab

IQR = Interquartile Range

Table 2.

Injury characteristics. Values are reported as counts and means (standard deviations) unless otherwise noted. Percentages are relative to the number of injuries recorded in the group summarized in each respective column. Previous hamstring strain injuries are reported if a prior timeloss injury occurred on the same limb of the injury included in the present study.

Reinjury
All (n = 26) Yes (n = 5) No (n = 17)
Previous Hamstring Strain Injury
 Yes 5 (19%) 0 (0%) 5 (24%)
 No 21 (81%) 5 (100%) 16 (76%)
Primary Muscle Injured
 Biceps Femoris Short Head 1 (4%) 0 (0%) 1 (5%)
 Biceps Femoris Long Head 19 (73%) 4 (80%) 15 (71%)
 Semitendinosus 2 (8%) 1 (20%) 1 (5%)
 Semimembranosus 4 (15%) 0 (0%) 4 (19%)
British Athletic Muscle Injury Classification
 1a 3 (12%) 2 (40%) 1 (5%)
 1b 3 (12%) 0 (0%) 3 (14%)
 2a 4 (15%) 0 (0%) 4 (19%)
 2b 2 (8%) 2 (40%) 0 (0%)
 2c 3 (12%) 1 (20%) 2 (10%)
 3a 1 (4%) 0 (0%) 1 (5%)
 3b 2 (8%) 0 (0%) 2 (10%)
 3c 8 (31%) 0 (0%) 8 (38%)
Muscle Edema (cm3/(kg*m)) 0.46 (0.54) 0.32 (0.15) 0.50 (0.59)
Fractional Anisotropy [-] 0.190(0.024) 0.206 (0.019) 0.187 (0.024)
Mean Diffusivity (mm2/s*103) 1.87 (0.15) 1.77 (0.05) 1.90 (0.16)
Radial Diffusivity (mm2/s*103) 1.69 (0.15) 1.58 (0.04) 1.71 (0.15)
λ1 (mm2/s*103) 2.25 (0.17) 2.16 (0.08) 2.27 (0.18)
λ2 (mm2/s*103) 1.82 (0.15) 1.72 (0.07) 1.85 (0.16)
λ3 (mm2/s*103) 1.55 (0.14) 1.45 (0.05) 1.58 (0.15)
Open in a new tab

Figure 3.

Figure 3.

Open in a new tab

Representative imaging data from one participant at time of injury, A.) T2-weighted, B.) mean diffusivity, and C.) primary eigenvalue maps along principal direction (λ1).

Time to Return to Sport

A significant association was not detected between days to RTS and any imaging measure (normalized muscle edema volume, IRR (95% CI) = 1.04 (0.76, 1.38), p = 0.81; BAMIC grade and site, p-values > 0.15; DTI metrics, p-values > 0.11) (Table 3).

Table 3.

Negative binomial regression models to determine the association between imaging parameters and return to sport (RTS) days. British Athletic Muscle Injury Classification (BAMIC) system were considered separately as overall as BAMIC grade (0-4) and BAMIC anatomical site (myofascial [a], musculotendinous [b], intratendinous [c]) and included as main effects in the model. Unit represents the unit increase used for interpretation of incident risk ratio (IRR) and 95% confidence interval (CI).

Model Parameter Unit IRR 95% CI p-Value
1 Muscle Edema (cm3/(m*kg)) 1 1.04 0.76 1.38 0.81
2 BAMIC Grade 1 1.22 0.85 1.75 0.29
BAMIC Site b (reference a) 1 2.09 1.24 3.53 0.15‡
BAMIC Site c (reference a) 1 1.48 0.81 2.73
3 Fractional Anisotropy [-] 0.01 1.06 0.99 1.13 0.11
4 Mean Diffusivity (mm2/s) 0.0001 0.92 0.79 1.06 0.25
5 Radial Diffusivity (mm2/s) 0.0001 0.91 0.78 1.05 0.19
6 λ1 (mm2/s) 0.0001 0.94 0.82 1.08 0.40
7 λ2 (mm2/s) 0.0001 0.91 0.78 1.05 0.18
8 λ3 (mm2/s) 0.0001 0.92 0.80 1.05 0.22
Open in a new tab
‡

Type III p-values based on score test statistic for categorical variable.

Reinjury

No significant associations were detected between normalized muscle edema volume (OR (95% CI) = 0.43 (0.14, 1.29), p = 0.13), BAMIC injury grade and site (p-values > 0.16), or FA values (OR = 1.57 (0.78, 3.19), p = 0.21) and reinjury. A significant association was detected between all diffusivity metrics except λ1 and reinjury, with a 1 unit (0.0001 mm/s2) increase in each of the diffusivity metrics at TOI resulting in the following decreased odds of reinjury: MD (OR = 0.46 (0.24,0.87), p = 0.016), RD (OR = 0.43 (0.23,0.81), p = 0.008), λ2 (OR = 0.44 (0.24,0.79), p = 0.007), λ3 (OR = 0.43 (0.23,0.82), p = 0.01). A near-significant relationship between λ1 and reinjury was detected (OR = 0.53 (0.27,1.02), p = 0.057) (Table 4, Figure 4).

Table 4.

Generalized estimating equation models to determine the relationship of imaging parameters between those that do versus do not go on to reinjure following hamstring strain injury. British Athletic Muscle Injury Classification (BAMIC) system were considered separately as BAMIC grade (0-4) and BAMIC anatomical site (myofascial [a], musculotendinous [b], intratendinous [c]) and included as main effects in the model. Unit represents the unit increase used for interpretation of odds ratio (OR) and 95% confidence interval (CI).

Model Parameter Unit OR 95% CI p-Value
1 † Muscle Edema (cm3/(m*kg)) 1 0.43 0.14 1.29 0.13
2 BAMIC Grade 1 0.33 0.07 1.55 0.16
BAMIC Site b (reference a) 1 1.19 0.09 15.80 0.99‡
BAMIC Site c (reference a) 1 0.95 0.06 14.97
3 Fractional Anisotropy [-] 0.01 1.57 0.78 3.19 0.21
4 Mean Diffusivity (mm2/s) 0.0001 0.46 0.24 0.87 0.016
5 Radial Diffusivity (mm2/s) 0.0001 0.43 0.23 0.81 0.008
6 λ1 (mm2/s) 0.0001 0.53 0.27 1.02 0.057
7 λ2 (mm2/s) 0.0001 0.44 0.24 0.79 0.007
8 λ3 (mm2/s) 0.0001 0.43 0.23 0.82 0.011
Open in a new tab
†

Model 1 used an independent correlation structure due to the lack of convergence when using an exchangeable correlation structure.

‡

Type III p-values based on score test statistic for categorical variable.

Figure 4.

Figure 4.

Open in a new tab

Generalized estimating equation model results demonstrating significant relationships with diffusivity metrics A.) mean diffusivity, B.) radial diffusivity, and principal effective diffusivity eigenvalue C.) λ2 and D.) λ3 in the injured region following acute hamstring strain injury and probability of reinjury.

Discussion

The purpose of this study was to determine the association of MRI-based injury grading and quantitative imaging measures at the time of acute HSI with time to RTS and reinjury. While Significant associations were not detected between any imaging measures and days to RTS, greater diffusivity measures (MD, RD, λ1, λ2, λ3) were associated with decreased odds of reinjury.

Time to Return to Sport

We did not detect an association between any of the imaging measures (edema volume, BAMIC injury grading, or quantitative parameters of muscle microstructure) and days to RTS. Although edema volume has been previously investigated, methods used to represent edema volume vary and are often estimated based on a representative slice determined as the greatest extent of injury (Crema et al., 2018; Reurink et al., 2015). To our knowledge, this is the first study that has measured muscle edema volume directly. Despite the elimination of variability in estimated volumes, we did not detect an association between edema volume and days to RTS. Because using DTI metrics to represent muscle microstructure is a relatively novel application, only one study to date has investigated DTI metrics and time to RTS following a variety of lower extremity muscle injuries and similarly did not find an association between the two (Biglands et al., 2020). Despite narrowing the inclusion criteria in the present study to compare across injuries of the hamstrings, we did not detect an association between DTI metrics and time to RTS.

The evidence supporting the association between BAMIC and injury prognosis is mixed. While our findings are consistent with many prior studies that did not find a relationship between BAMIC and days to RTS (Shamji et al., 2021; Wangensteen et al., 2015; Wangensteen et al., 2018), other studies have demonstrated a significant relationship does exist (Biglands et al., 2020; McAuley et al., 2022; Pollock et al., 2016; Tears et al., 2022). BAMIC injury classification consists of two components, a grade (0-4) based on the relative amount of hyperintense signal change and a sub-classification based on the anatomical site (a-c) of injury (Pollock et al., 2016). Inconsistencies in how each component is used in statistical models to determine clinical associations may contribute to the mixed findings regarding this tool.

Further, differences in study designs and external factors influencing time to RTS may also explain inconsistencies in relationships between TOI imaging measures and clinical outcomes. For example, the inclusion of athletes with and without MRI-confirmed injuries complicates comparisons across studies, as distinct relationships between imaging findings and time to RTS have been identified with athletes that do (MRI-positive) versus do not (MRI-negative) demonstrate injury on imaging (Gibbs et al., 2004; Kumaravel et al., 2018; Reurink et al., 2015). Inclusion of athletes with only MRI-positive injuries in the present study was necessary for the identification of all imaging parameters. However, limiting the subject pool to those with only MRI-positive injuries may have contributed to the lack of identified relationships. Finally, ranges of time to RTS often vary greatly within and across studies, potentially due to the differences in study populations and the influence of external factors such as RTS criteria, timing relative to the competitive season, and exposure after RTS. While imaging measures at TOI may still hold relevant relationships with long-term outcomes following HSI, future studies will need a rigorous study design to account for external factors and challenges associated with using time to RTS as an outcome measure. Consideration of interval assessments of recovery independent of time to RTS or other measures of longitudinal outcomes such as resolution of strength or performance may be warranted.

Reinjury

Of the 26 index HSIs, 5 (19%) went on to reinjure within 12 months. Consistent with a recent review (Green et al., 2020), we did not detect an association between TOI MRI-based injury assessments at the gross anatomical level and reinjury. Conversely, DTI measures of muscle microstructure indicate that decreased diffusivity measures (MD, RD, λ1, λ2, λ3) were associated with greater odds of reinjury.

In the presence of muscle injury, less restricted diffusion is expected due to compromise of structural integrity of the sarcolemma, which affects the permeability of water exchange between intra- and extracellular compartments. This would be consistent with changes in diffusivity parameters including increased diffusivity in principal eigenvalues (λ1, λ2, λ3), increased MD, RD, and decreased microstructural organization (FA). Preliminary evidence indicates these trends exist following acute muscle strain injuries (Biglands et al., 2020; Giraudo et al., 2018; Monte et al., 2023; Zaraiskaya et al., 2006). Given that increased diffusivity measures are consistent with the physiological processes expected with an acute muscle strain injury, it is probable that a more severe injury would be associated with increased measures of diffusivity and decreased values of tissue organization (FA) and thus more likely to reinjure. However, the opposite was true in our study, with those who reinjured demonstrating a decrease in diffusivity measures (MD, RD, λ1, λ2, λ3) compared to those who do not reinjure. Findings from the current study demonstrating that those who go on to reinjure have muscle microstructural metrics more typical of non-injured tissue at the time of the index injury may be a result of external factors not captured by imaging. Although the relationship of diffusion changes in the presence of injury is complex (Berry et al., 2018), it is possible that findings indicating decreased diffusivity at the time of injury may reflect the degree of the elicited healing response which in turn may be associated with reinjury. Further, it may also be possible that factors such as the state of muscle microstructure at RTS or the relative amount of recovery compared to TOI may play a role in the directional relationship observed in this study with decreased diffusivity metrics at the TOI in athletes who went on to reinjure.

Limitations and Future Work

The validity of DTI metrics used to represent muscle microstructure has been established (Napadow et al., 2001; Van Donkelaar et al., 1999). However, the clinical utility of these measures has not been extensively studied. To our knowledge this is one of the first studies that relates quantitative DTI at TOI with clinical outcomes such as time to RTS and reinjury. While we acknowledge that limitations exist in the present study, findings from this study may help to contribute to further research on this topic. Although the focus of the present analysis was to focus on the relationships between quantitative MRI-based characteristics of injury and clinical outcomes, HSIs are known to be multifactorial (Green et al., 2020). The investigation of potential interactions between imaging measures and additional risk factors such as prior injury and exposure after RTS is warranted. While the reinjury rate observed in the present study is consistent with prior prospective studies (Biglands et al., 2020; McAuley et al., 2022), accounting for an extensive list of potential contributing factors to explain reinjury following HSI will require a larger sample with adequate power. An appropriate sample size for such study could be estimated using the ratio of one confounding variable for every ten recorded events (Peduzzi et al., 1996), or in this case hamstring reinjuries. For example, in addition to at least one imaging measure as an explanatory variable, accounting for additional confounding variables such as previous hamstring injury, the sport played by the injured athlete, and whether or not the rehabilitation and thus RTS determination was made in- or out-of-season would require upwards of 40 reinjuries and a minimum of 210 index injuries, assuming a 19% reinjury rate as observed in the present study.

Furthermore, although the present study design was prospective, several additional factors must be considered when creating a clinical prediction model including appropriate sample size calculation, extensive testing of the created prediction model (internal and external validation), considerations for missing data, and a multicenter study to improve generalizability (Bullock et al., 2021). In addition, future investigations may consider relative changes in imaging measures throughout recovery and at the time of RTS to fully define the utility of quantitative MRI following HSI.

Conclusions

An association was not detected between MRI-based measures of injury classification (BAMIC) and muscle edema volume assessed at time of HSI with days to RTS or reinjury. Similarly, a significant association between quantitative diffusion parameters at TOI and days to RTS was not detected. However, those who went on to reinjure demonstrated decreased quantitative diffusion parameters within the injured muscle at TOI compared to those who do not reinjure. Despite recent advancements in the ability of MRI to measure muscle microstructure through DTI, the relationship between TOI imaging assessment and longitudinal clinical outcomes remains complicated. While DTI following HSI may still hold potential implications for health and sports professionals, additional work is warranted to further understand the relationship between DTI metrics and reinjury status and the influence of external factors on the relationship between TOI imaging measures and longitudinal clinical outcomes.

Acknowledgements

The authors would like to acknowledge the Sports Medicine staff in the University of Wisconsin-Madison Division of Athletics for their commitment to the welfare of the student athletes and contributions to the Badger Athletic Performance program.

Funding

This work is supported by NBA & GE HealthCare Orthopedics and Sports Medicine Collaboration (MYT-015, D223) and NIH award TL1TR002375. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Conflict of Interest

Select authors of this manuscript declare relationships with the following companies: Dr. Bryan Heiderscheit declares a potential conflict of interest directly related to this work (research support to institution from NBA and GE Healthcare). The authors (CMW, SAH, MRSJ, KL, RK) of this manuscript declare no direct relationships with any companies, whose products or services may be related to the subject matter of the article.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Andersson JL, Skare S, Ashburner J, 2003. How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging. Neuroimage 20, 870–888. [DOI] [PubMed] [Google Scholar]
  2. Andersson JLR, Sotiropoulos SN, 2016. An integrated approach to correction for off-resonance effects and subject movement in diffusion MR imaging. Neuroimage 125, 1063–1078. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Berry DB, Regner B, Galinsky V, Ward SR, Frank LR, 2018. Relationships between tissue microstructure and the diffusion tensor in simulated skeletal muscle. Magn Reson Med 80, 317–329. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Biglands JD, Grainger AJ, Robinson P, Tanner SF, Tan AL, Feiweier T, Evans R, Emery P, O’Connor P, 2020. MRI in acute muscle tears in athletes: can quantitative T2 and DTI predict return to play better than visual assessment? Eur Radiol 30, 6603–6613. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bullock GS, Hughes T, Sergeant JC, Callaghan MJ, Riley RD, Collins GS, 2021. Clinical Prediction Models in Sports Medicine: A Guide for Clinicians and Researchers. J Orthop Sports Phys Ther 51, 517–525. [DOI] [PubMed] [Google Scholar]
  6. Crema MD, Godoy IRB, Abdalla RJ, de Aquino JS, Ingham SJM, Skaf AY, 2018. Hamstring Injuries in Professional Soccer Players: Extent of MRI-Detected Edema and the Time to Return to Play. Sports Health 10, 75–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Damon BM, Froeling M, Buck AK, Oudeman J, Ding Z, Nederveen AJ, Bush EC, Strijkers GJ, 2017. Skeletal muscle diffusion tensor-MRI fiber tracking: rationale, data acquisition and analysis methods, applications and future directions. NMR Biomed 30. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. de Visser HM, Reijman M, Heijboer MP, Bos PK, 2012. Risk factors of recurrent hamstring injuries: a systematic review. Br J Sports Med 46, 124–130. [DOI] [PubMed] [Google Scholar]
  9. Gibbs NJ, Cross TM, Cameron M, Houang MT, 2004. The accuracy of MRI in predicting recovery and recurrence of acute grade one hamstring muscle strains within the same season in Australian Rules football players. J Sci Med Sport 7, 248–258. [DOI] [PubMed] [Google Scholar]
  10. Giraudo C, Motyka S, Weber M, Karner M, Resinger C, Feiweier T, Trattnig S, Bogner W, 2018. Normalized STEAM-based diffusion tensor imaging provides a robust assessment of muscle tears in football players: preliminary results of a new approach to evaluate muscle injuries. Eur Radiol 28, 2882–2889. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Green B, Bourne MN, van Dyk N, Pizzari T, 2020. Recalibrating the risk of hamstring strain injury (HSI): A 2020 systematic review and meta-analysis of risk factors for index and recurrent hamstring strain injury in sport. Br J Sports Med 54, 1081–1088. [DOI] [PubMed] [Google Scholar]
  12. Greenky M, Cohen SB, 2017. Magnetic resonance imaging for assessing hamstring injuries: clinical benefits and pitfalls - a review of the current literature. Open Access J Sports Med 8, 167–170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Guermazi A, Roemer FW, Robinson P, Tol JL, Regatte RR, Crema MD, 2017.Imaging of Muscle Injuries in Sports Medicine: Sports Imaging Series. Radiology 285, 1063. [DOI] [PubMed] [Google Scholar]
  14. Handsfield GG, Meyer CH, Hart JM, Abel MF, Blemker SS, 2014. Relationships of 35 lower limb muscles to height and body mass quantified using MRI. J Biomech 47, 631–638. [DOI] [PubMed] [Google Scholar]
  15. Heiderscheit BC, Sherry MA, Silder A, Chumanov ES, Thelen DG, 2010. Hamstring strain injuries: recommendations for diagnosis, rehabilitation, and injury prevention. J Orthop Sports Phys Ther 40, 67–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hickey J, Shield AJ, Williams MD, Opar DA, 2014. The financial cost of hamstring strain injuries in the Australian Football League. Br J Sports Med 48, 729–730. [DOI] [PubMed] [Google Scholar]
  17. Jacobsen P, Witvrouw E, Muxart P, Tol JL, Whiteley R, 2016. A combination of initial and follow-up physiotherapist examination predicts physician-determined time to return to play after hamstring injury, with no added value of MRI. Br J Sports Med 50, 431–439. [DOI] [PubMed] [Google Scholar]
  18. Jenkinson M, Beckmann CF, Behrens TE, Woolrich MW, Smith SM, 2012. Fsl. Neuroimage 62, 782–790. [DOI] [PubMed] [Google Scholar]
  19. Kumaravel M, Bawa P, Murai N, 2018. Magnetic resonance imaging of muscle injury in elite American football players: Predictors for return to play and performance. Eur J Radiol 108, 155–164. [DOI] [PubMed] [Google Scholar]
  20. LIu H GW, Moorman CT, Yu B , 2012. Injury rate, mechanism, and risk factors of hamstring strain injuries in sports: A review of the literature. Journal of Sport and Health Science 1, 92–101. [Google Scholar]
  21. Manjon JV, Coupe P, Concha L, Buades A, Collins DL, Robles M, 2013. Diffusion weighted image denoising using overcomplete local PCA. PLoS One 8, e73021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. McAleer S, Macdonald B, Lee J, Zhu W, Giakoumis M, Maric T, Kelly S, Brown J, Pollock N, 2022. Time to return to full training and recurrence of rectus femoris injuries in elite track and field athletes 2010-2019; a 9-year study using the British Athletics Muscle Injury Classification. Scand J Med Sci Sports 32, 1109–1118. [DOI] [PubMed] [Google Scholar]
  23. McAuley S, Dobbin N, Morgan C, Goodwin PC, 2022. Predictors of time to return to play and re-injury following hamstring injury with and without intramuscular tendon involvement in adult professional footballers: A retrospective cohort study. J Sci Med Sport 25, 216–221. [DOI] [PubMed] [Google Scholar]
  24. McCarthy P, 2022. FSLeyes. Zenodo. [Google Scholar]
  25. Monte JR, Hooijmans MT, Froeling M, Oudeman J, Tol JL, Strijkers GJ, Nederveen AJ, Maas M, 2023. Diffusion tensor imaging and quantitative T2 mapping to monitor muscle recovery following hamstring injury. NMR Biomed, e4902. [DOI] [PubMed] [Google Scholar]
  26. Napadow VJ, Chen Q, Mai V, So PT, Gilbert RJ, 2001. Quantitative analysis of three-dimensional-resolved fiber architecture in heterogeneous skeletal muscle tissue using nmr and optical imaging methods. Biophys J 80, 2968–2975. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Opar DA, Williams MD, Shield AJ, 2012. Hamstring strain injuries: factors that lead to injury and re-injury. Sports Med 42, 209–226. [DOI] [PubMed] [Google Scholar]
  28. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR, 1996. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 49, 1373–1379. [DOI] [PubMed] [Google Scholar]
  29. Pollock N, James SL, Lee JC, Chakraverty R, 2014. British athletics muscle injury classification: a new grading system. Br J Sports Med 48, 1347–1351. [DOI] [PubMed] [Google Scholar]
  30. Pollock N, Patel A, Chakraverty J, Suokas A, James SL, Chakraverty R, 2016. Time to return to full training is delayed and recurrence rate is higher in intratendinous (‘c’) acute hamstring injury in elite track and field athletes: clinical application of the British Athletics Muscle Injury Classification. Br J Sports Med 50, 305–310. [DOI] [PubMed] [Google Scholar]
  31. Reurink G, Brilman EG, de Vos RJ, Maas M, Moen MH, Weir A, Goudswaard GJ, Tol JL, 2015. Magnetic resonance imaging in acute hamstring injury: can we provide a return to play prognosis? Sports Med 45, 133–146. [DOI] [PubMed] [Google Scholar]
  32. Reurink G, Goudswaard GJ, Tol JL, Almusa E, Moen MH, Weir A, Verhaar JA, Hamilton B, Maas M, 2014. MRI observations at return to play of clinically recovered hamstring injuries. Br J Sports Med 48, 1370–1376. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Rudisill SS, Kucharik MP, Varady NH, Martin SD, 2021. Evidence-Based Management and Factors Associated With Return to Play After Acute Hamstring Injury in Athletes: A Systematic Review. Orthop J Sports Med 9, 23259671211053833. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Shamji R, James SLJ, Botchu R, Khurniawan KA, Bhogal G, Rushton A, 2021. Association of the British Athletic Muscle Injury Classification and anatomic location with return to full training and reinjury following hamstring injury in elite football. BMJ Open Sport Exerc Med 7, e001010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Sherry MA, Johnston TS, Heiderscheit BC, 2015. Rehabilitation of acute hamstring strain injuries. Clin Sports Med 34, 263–284. [DOI] [PubMed] [Google Scholar]
  36. Tears C, Rae G, Hide G, Sinha R, Franklin J, Brand P, Hasan F, Chesterton P, 2022. The British Athletics Muscle Injury Classification grading system as a predictor of return to play following hamstrings injury in professional football players. Phys Ther Sport 58, 46–51. [DOI] [PubMed] [Google Scholar]
  37. Van Donkelaar CC, Kretzers LJ, Bovendeerd PH, Lataster LM, Nicolay K, Janssen JD, Drost MR, 1999. Diffusion tensor imaging in biomechanical studies of skeletal muscle function. J Anat 194 ( Pt 1), 79–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. van Heumen M, Tol JL, de Vos RJ, Moen MH, Weir A, Orchard J, Reurink G, 2017. The prognostic value of MRI in determining reinjury risk following acute hamstring injury: a systematic review. Br J Sports Med 51, 1355–1363. [DOI] [PubMed] [Google Scholar]
  39. Wangensteen A, Almusa E, Boukarroum S, Farooq A, Hamilton B, Whiteley R, Bahr R, Tol JL, 2015. MRI does not add value over and above patient history and clinical examination in predicting time to return to sport after acute hamstring injuries: a prospective cohort of 180 male athletes. Br J Sports Med 49, 1579–1587. [DOI] [PubMed] [Google Scholar]
  40. Wangensteen A, Guermazi A, Tol JL, Roemer FW, Hamilton B, Alonso JM, Whiteley R, Bahr R, 2018. New MRI muscle classification systems and associations with return to sport after acute hamstring injuries: a prospective study. Eur Radiol 28, 3532–3541. [DOI] [PubMed] [Google Scholar]
  41. Zaraiskaya T, Kumbhare D, Noseworthy MD, 2006. Diffusion tensor imaging in evaluation of human skeletal muscle injury. J Magn Reson Imaging 24, 402–408. [DOI] [PubMed] [Google Scholar]

RESOURCES